Materials and methods

Patients and BCC samples

Material of 13 BCC was available from five psoriasis patients (A-E) who had a history of PUVA treatment. The exact patients' demographics, risk factors, PUVA data, and tumor characteristics are given in Table I.

Table I - Patients' demographics, risk factors, PUVA data, and tumor characteristics.

Full table

DNA extraction

For DNA extraction paraffin-embedded tumor samples were dissected to eliminate normal tissue adjacent to and in the tumors to reduce the portion of nontumor cells in the samples. Briefly, 7-10 m tissue sections were deparaffinized with xylene (10 min) and ethanol (10 min), air dried, and then damped into deionized water. The dissected cells were then suspended in 50 l of a solution containing 0.1 M Tris-HCl, pH 8.0, and 1 g per l proteinase K; incubated in a water bath at 55°C overnight; and then boiled for 10 min. The preparation was stored at -20°C until used.

PCR-SSCP analysis

DNA from exons 4–9 of the p53 gene was amplified using the primer sequences as listed in Table II. Primers with slight modifications were used to amplify exons 4 and 5 (^{Brash et al, 1991}) and exons 6–9 (^{Wang et al, 1997}). Four microliters of each lysate was used as the template in a 50 l solution containing 10 mM Tris-HCl pH 8.3, 50 mM KCl, 1.5 mM deoxyribonucleoside triphosphate at 100 M each, upstream and downstream primers 10 pM each, and 2.5 U AmpliTag Gold (Perkin Elmer, Vienna, Austria). PCR assays were run for 40 cycles of amplification as follows: denaturation at 94°C for 45 s, annealing at 60°C for 30 s, and extension at 72°C for 30 s. Before the first cycle, tubes were incubated for 12 min at 94°C, after the last cycle, for 7 min at 72°C. To avoid potential contamination of PCR and to rule out PCR-generated mutations, every PCR reaction included blank samples lacking DNA templates and other samples containing human placental DNA. After amplification, 50 l of the PCR products was fractionated by electrophoresis in 2.5% MetaPhor agarose gel (FMC Bioproducts, Rockland, ME). Amplified and purified fragments were cut out of the gel and extracted using a commercially available gel extraction kit (Qiagen, Valencia, CA).

Table II - Primers used for PCR-SSCP and sequencing analysis of p53.

Full table

For SSCP, 0.5 l (30 ng) amplified and purified DNA, 2.5 l sample buffer, and 3 l denaturing solution were denaturated at 95°C for 10 min and thereafter placed on ice. Samples were loaded on to mutation detection enhancement gels (Gene Gel Excel 12.5/24, Amersham Pharmacia Biotech, Vienna, Austria). The gels were then run for 95 min at 10°C (exons 7–9), 15°C (exon 6), and 20°C (exons 4 and 5). Running conditions were as follows: 600 V, 25 mA, and 15 W. Once run, gels were stained using a DNA Silver Staining Kit (Amersham Pharmacia Biotech). At least two independent PCR–SSCP analyses were carried out per tumor sample.

Nucleotide sequencing

All amplified and purified samples were: (i) sequenced using 3.2 pM primer, 30 ng DNA, and sequencing reagents (Cycle Sequencing Ready Reaction; Perkin Elmer, Foster City, CA); (ii) precipitated; and (iii) separated in an ABI Prism 310 system (Perkin Elmer). Both DNA strands were sequenced, and mutations on one strand were always confirmed on the opposite strand. Sequence analysis was done on a Power Macintosh G3 (Apple Computer, Cupertino, CA) using Sequence Analysis Software (Perkin Elmer).

Statistical analysis

Differences in the number of mutations types between different groups of subjects were analyzed for statistical significance using Fisher's exact test in the StatView statistical analysis program version 5.0.1 (SAS Institute, Cary, NC). p 0.05 was considered significant.

Results

Both UV- and PUVA-type p53 mutations are present in BCC from PUVA-treated patients

A total of 13 BCC were obtained from five PUVA-treated psoriasis patients and analyzed for p53 mutations by PCR–SSCP analysis and sequencing of exons 4–9. Sequencing revealed a total of 17 mutations (11 mis-sense, two non-sense, and four silent mutations) in seven of 13 BCC (54%) Table III. Mutations of the UV type [CCTT transition Figure 1a and CT transition Figure 1c at dipyrimidine sites] and the PUVA type (TC transition at a 5'-TpA site) Figure 1b were seen. Of the 13 mis-sense or non-sense mutations, 12 (92%) were at dipyrimidine sites and nine (69%) were of the UV type (eight CT transitions and one CCTT transition). Three of the CT transitions were at dipyrimidine sites opposite a 5'-TpG sequence (a potential psoralen-binding site) (A1, codon 94; A3, codon 67; and D1, codon 317) Table III. As this left the origin of these mutations (UV or PUVA) ambiguous, they were considered separately Figure 2. In addition, two other mutations (15%) occurred at 5'-TpG sites (including one at a dipyrimidine and one at a nondipyrimidine site), one mutation (8%) at a 5'-TpA site (all of potential PUVA origin), and one involved a (G:CT:A) transversion (8%). In addition, a BCC from patient C (tumor C1) showed a GC transversion in exon 4 of codon 72, a known p53 polymorphism (^{Olschwang et al, 1991}). Two tumors (A1 and A3) had multiple mis-sense mutations (most of them of UV type) at exons 4 and 5, respectively Table III. In support of a previous study (^{Nataraj et al, 1997}), the greater intensity of wild-type vs shifted (i.e, mutant) bands in the PCR–SSCP analysis (data not shown) suggested that: (i) in most cases only one p53 allele was mutated and the other was wild-type; (ii) only a portion of tumor cells harbored mutations; and/or (iii) contaminating normal cells were present in the tumor samples. Importantly, this is also supported by the observation that most of our electropherograms simultaneously showed wild-type and mutated DNA sequences Figure 1, consistent with the SSCP results. Taken together, these findings suggested that there was no loss of heterozygosity at the p53 gene of the BCC. Importantly, however, all p53 mutations detected were somatic because they were not present in normal tissue samples from the patients.

Figure 1.

Sequence analysis reveals UV- and PUVA-type p53 mutations in PUVA-associated BCC. The upper panel shows electropherograms of DNA from (A)tumor A4(exon 5, codon 177), CCTT transition of UV type; (B) tumor E3 (exon 7, codon 234), TC transition on a potential psoralen binding 5'-TpA:ApT site (i.e, of PUVA type); and (C)tumor A2 (exon 5, codon 155), CT transition of UV type. The lower panel shows electropherograms of wild-type DNA from placenta.

Full figure and legend (38K)

Figure 2.

BCC from PUVA-treated patients have a higher proportion of mutations at 5'-TpG sites (i.e, mutations at potential psoralen-binding sites and thus potentially of PUVA origin) than do BCC in the general population. p53 mutations detected in BCC from the general population in several previous studies (for details see Materials and Methods) were pooled and compared with those in BCC from PUVA-treated patients in terms of type and number. *Note that there were three CT transitions at dipyrimidine sites, each of which occurred opposite a 5'-TpG sequence (a potential psoralen-binding site) (see Table III). In this graph, these mutations are represented at the top of the column marked CT and at the top of the column marked 5'-TpG at dipyrimidine sites. The origin of these mutations (UV or PUVA) was ambiguous, and thus they were considered separately in the mutation spectrum analysis. ⁺The proportion of p53 mutations at 5'-TpG sites was statistically significantly higher in BCC from PUVA-treated patients than in those from the general population [38% (five of 13) vs 2% (two of 92), respectively; Fisher's exact test, p 0.001].

Full figure and legend (23K)

Table III - Type of p53 mutations detected in BCC from PUVA-treated psoriasis patients.

Full table

BCC from PUVA-treated patients have a similar incidence of UV-type mutations but a higher proportion of PUVA-type mutations than BCC from the general population

Data on the type and number of p53 mutations in BCC from non-PUVA-treated subjects of the general population were gathered from several previous studies (^{Rady et al, 1992;Moles et al, 1993;Sato et al, 1993;Ziegler et al, 1993;Kanjial et al, 1995;Gailani et al, 1996;D'Errico et al, 1997;Ponten et al, 1997}), pooled, and compared with the data on the mutations found in BCC from PUVA-treated patients Figure 2. The total percentage of mis-sense or non-sense UV fingerprint mutations at dipyrimidine sites (i.e, CT and CCTT transitions) was 69% (nine of 13) [or 46% (six of 13), when not counting the three CT mutations at 5'-TpG sites] in PUVA-treated patients and 64% (59/92) in the general population. The difference was deemed not statistically significant when subjected to Fisher's exact test (p 0.2); however, the percentage of p53 mutations at 5'-TpG sites in BCC from PUVA-treated patients was statistically significantly higher than that in the general population [38% (five of 13) vs 2% (two of 92); Fisher's exact test, p 0.001].

Discussion

In this study, we examined the hypothesis that carcinogen-specific mutations are present in the p53 gene of BCC from PUVA-treated patients. We found a total of 13 mis-sense or non-sense p53 mutations in seven of 13 BCC (54%) and that 12 of these 13 mutations (92%) were at dipyrimidine sites and that nine (69%) of the mutations had the UV fingerprint (CT or CCTT transitions) (^{Brash et al, 1991}). More important, the percentage of UV fingerprints in PUVA-associated BCC did not significantly differ from that in the general population Figure 2. There are several possible explanations for this high proportion of UV fingerprint p53 mutations. First, most BCC examined were from heavily sun-exposed body areas; on the other hand no PUVA-associated BCC from other body sites were available for study. Second, all patients in this study had a history of UVB treatment (a possible cause of UV-type p53 mutations). Third, fluorescent tubes used for PUVA therapy in Europe emit a small portion of radiation within the UVB range (^{Nataraj et al, 1997}), which may in turn cause UV-type mutations. Fourth, stable oxidation products of cytosine and thymidine induced by 8-MOP and UVA or by UVA alone could also induce CT transitions (^{Wang et al, 1998}). Finally, PUVA generates reactive oxygen species, which also may induce CCTT transitions (^{Reid and Loeb, 1993}). The hypothesis that exposure to PUVA or UVA alone may result in UV-type mutations, however, is strongly contradicted by results of cell culture (^{Sage et al, 1993;Chiou and Yang, 1995;Gunther et al, 1995}) and murine (^{Nataraj et al, 1996}) studies, in which UV-type mutations were rarely if ever observed after PUVA exposure.

An important finding in this study was that six of the 13 p53 mutations identified (46%) were at potential psoralen-binding sites and thus may have been induced by PUVA; however, only one mutation (8%) was at a 5'-TpA site compared with five (38%) at 5'-TpG sites. This finding is similar to the results of a previous study (^{Nataraj et al, 1997}) on PUVA-associated SCC, in which 13 of 25 p53 mutations (52%) were found at 5'-TpG sites. Although cell culture (^{Sage et al, 1993;Chiou and Yang, 1995;Gunther et al, 1995}) and murine studies (^{Nataraj et al, 1996}) revealed that 5'-TpA sites are the most common targets for PUVA mutagenesis, a cell culture study (^{Gunther et al, 1995}) using low (clinically relevant) doses of PUVA revealed that the highest percentage (approximately one-third) of mutations in murine fibroblasts containing supF DNA occurred at 5'-TpG sites. Interestingly, in a study on PUVA-associated SCC our group of investigators found frequent mutation of the Ha-ras oncogene and a 28% incidence of mis-sense/non-sense mutations (five of 18) at 5'-TpG sites (^{Kreimer-Erlacher et al, in press}). An alternative explanation for PUVA-induced mutations at 5'-TpG sites was recently proposed namely, reactive oxygen species-mediated damage at G, C, and/or T sites (^{Wang et al, 1998;Peritz and Gasparro, 1999}).

Because molecular epidemiology studies can link mutations to causative agents, they are a potent tool for determining the factors that contribute to cancers (^{Peritz and Gasparro, 1999}). In the case of PUVA-associated skin cancer, however, the situation is made very complex by the concomitant presence of multiple risk factors such as exposure to potentially carcinogenic and cocarcinogenic agents including UVB, X-ray radiation, medical tar, and methotrexate (^{Maier et al, 1996}). For instance, two patients in the present study (C and E) had been exposed in the past to therapeutic arsenic, a well-known carcinogen that causes skin cancer after long latency periods. We found two PUVA-type mis-sense mutations (one at a 5'-TpA site and one at 5'-TpG site) in the six BCC samples from those patients but no mutations of any other type; however, arsenic exposure may lead to GT transversions via the production of reactive oxygen species (^{Shibutani et al, 1991}). Interestingly, in another study, we had found two CT transitions and one GT transversion in the p53 gene of a total of four BCC from two arsenic-exposed, non-PUVA-treated patients (Seidl et al unpublished results). GT transversions, however, can also be caused by exposure to UV-induced singlet oxygen via production of 8-hydroxyguanosine (^{Cheng et al, 1992}), to benzo[a]pyrenes (e.g, from tobacco smoke) (^{Chiba et al, 1990}), or to hydrogen radicals generated by -radiation (^{Braun et al, 1993}). Patients A, C, and E in our study all had a history of medical treatment with tar, which may contain carcinogenic compounds, such as benzo[a]pyrenes (^{Godschalk et al, 1998}). Indeed, one tumor (A1) harbored at exon 5 (codon 144) a GT (G:CT:A) transversion. Patient A had also received X-ray radiation on the scalp and oral methotrexate for the treatment of psoriasis. It is well-known that X-ray treatment can cause BCC, possibly by causing GT mutations (^{Braun et al, 1993}).

Two BCC (A1 and A3) had multiple mis-sense mutations, mostly of the UV type Table III. As sequencing and PCR–SSCP were carried out on each exon separately and vector cloning was not performed in our study, it was therefore impossible to determine whether the mutations in the different exons of the tumors occurred in the same p53 allele or in different alleles. Interestingly, in this context, a previous study by^{Ziegler et al (1993)} revealed that 56% of human BCC contained p53 mutations and that 45% of the tumors contained a second point mutation on the other p53 allele. Consistent with the results of our study, however,^{D'Errico et al (1997)} previously reported that all p53 mutated BCC they studied apparently retained the wild-type p53 allele suggesting that only one p53 allele seems to be inactivated in BCC. Another previous study found multiple p53 mutations in eight of 12 PUVA-associated human SCC (^{Nataraj et al, 1997}). In a study of PUVA-induced SCC in mice, some tumors harbored multiple p53 mutations (^{Nataraj et al, 1996}). Indeed, the presence of multiple p53 mutations in cancers suggests that clones harboring an initial mutation on one allele are targets for a second mutation on the other allele or that these mutations may arise independently, perhaps in different clonal subpopulations during tumor development (^{Nataraj et al, 1996}). Indeed the latter hypothesis is supported by our finding that some BCC (A2, A3, and E3) had silent mutations (i.e, no change in amino acid sequence) of the UV or PUVA type Table III and that at least in some cases only a portion of tumor cells harbored p53 mutations Figure 1. Clearly, as was the case with the PUVA-associated SCC studied by^{Nataraj et al (1997)}, not all p53 mutations in tumors may contribute to tumor development as some mutations may arise after tumor initiation by repeated exposure to agents such as PUVA and/or UV(B).

Our present results are also important because they contradict the previous finding of^{Greenblatt et al (1994)} that approximately 90% of all p53 mutations in cancers occur between exons 5 and 8. In contrast we found that: (i) seven of 13 p53 mis-sense/non-sense mutations (54%) occurred in the nonconserved domain of exon 4 (codons 67–94); (ii) three (23%) in the conserved domain III of exon 5 (codons 144–177); (iii) two (15%) in the conserved domain IV of exon 7 (codons 229 and 234); and (iv) one (8%) in the nonconserved domain of exon 9 (codon 317). In previous studies, pooling of DNA sequencing results revealed that BCC appeared to have a major mutational hot spot between codons 241 and 280 and a minor region between codons 161 and 200 (^{Nataraj et al, 1995}; references cited therein). In this study, we observed two mis-sense mutations (i.e, CT transitions at a multipyrimidine site) at codon 88 of exon 4 (tumors A1 and A3), but nowhere else a hotspot mis-sense mutation incidence on the p53 gene. Mutations in conserved domains, of the p53 gene have been suggested to destabilize the p53 protein (^{Cho et al, 1994}), but little is known about mutations outside the conserved core domain.

In conclusion, our study revealed that the majority of p53 mutations in BCC from PUVA-treated patients occurred at dipyrimidine sites and had the UV fingerprint, similar to the case in the general population Figure 2; however, a smaller percentage of the mutations occurred at 5'-TpG sites, suggesting that even though the major cause of p53 mutations (and possibly BCC tumorigenesis) seems to be environmental and/or therapeutic UV(B) exposure, PUVA itself may directly cause some p53 mutations as well. Moreover, as only about half of the BCC in the present study harbored p53 mutations, it may be that other tumor suppressor genes and/or oncogenes are targets for PUVA-associated damage leading to BCC tumorigenesis. One particular candidate that needs to be investigated is the "patched" gene, which is frequently mutated in BCC of patients with the basal cell nevus syndrome and in the general population (reviewed in^{Wikonkal and Brash, 1999}). This study, however, highlights the value of the new field of molecular epidemiology as the results confirm on the molecular level those of many conventional epidemiologic PUVA follow-up studies, suggesting that PUVA itself may not be the major factor in BCC formation in PUVA-treated subjects (^{Stern et al, 1998}).

References

References

1.	Brash DE, Rudolph JA & Simon JA et al. A role for sunlight in skin cancer: UV-induced p53 mutations in squamous cell carcinoma. Proc Natl Acad Sci USA (1991) 90: 10124–10128.
2.	Braun JE, Wanamarta AH, van den Akker E, Lafleur MV & Retel J. C/G to A/T transversions represent the main type of mutation induced by -irradiation in double stranded M13mp10 DNA in a nitrogen-saturated solution. Mutat Res (1993) 289: 255–263. | PubMed | ISI | ChemPort |
3.	Bruynzeel I, Bergman W, Hartevelt HM, Kenter CC, Van de Velde EA, Schothorst AA & Suurmond D. "High single-dose" European PUVA regimen also causes an excess of non-melanoma skin cancer. Br J Dermatol (1991) 124: 49–55. | PubMed | ISI | ChemPort |
4.	Cheng KC, Cahill DS, Kasai H, Nishimura S & Loeb LA. 8-Hydroxyguanosine, an abundant form of oxidative DNA damage, causes G:T and A:C substitutions. J Biol Chem (1992) 267: 166–172. | PubMed | ISI | ChemPort |
5.	Chiba T, Takahashi T & Nau MM et al. Mutations in the p53 gene are frequent in primary, resected, non-small cell lung cancer. Oncogene (1990) 5: 1603–1610. | PubMed | ISI | ChemPort |
6.	Chiou C & Yang J. Mutagenicity and specific mutation spectrum induced by 8-methoxypsoralen plus a low dose of UVA in the hprt gene in diploid human fibroblasts. Carcinogen (1995) 16: 1357–1362. | ISI | ChemPort |
7.	Cho Y, Corina S, Jeffrey P & Pavlevitch N. Crystal structure of a p53 tumor suppressor-DNA complex: Understanding tumorigenic mutations. Science (1994) 265: 346–355. | PubMed | ISI | ChemPort |
8.	D'Errico M, Calcagnile AS & Corona R et al. p53 mutations and chromosome instability in basal cell carcinomas developed at an early or late age. Cancer Res (1997) 57: 747–752. | PubMed | ChemPort |
9.	Gailani MR, Leffell DJ, Ziegler AM, Gross EG, Brash DE & Bale AE. Relationship between sunlight exposure and a key genetic alteration in basal cell carcinoma. J Natl Cancer Inst (1996) 88: 349–354. | PubMed | ChemPort |
10.	Godschalk RW, Ostertag JU, Moonen EJ, Neumann HA, Kleinjans JC & van Schooten FJ. Aromatic DNA adducts in human white blood cells and skin after dermal application of coal tar. Cancer Epidemiol Biomarkers Prev (1998) 7: 767–773. | PubMed | ISI | ChemPort |
11.	Greenblatt MS, Bennett WP, Hollstein M & Harris CC. Mutations in the p53 tumor suppressor gene: Clues to cancer etiology and molecular pathogenesis. Cancer Res (1994) 54: 4855–4878. | PubMed | ISI | ChemPort |
12.	Gunther EJ, Yeasky TM, Gasparro FP & Glazer PM. Mutagenesis by 8-methoxypsoralen and 5-methylangelicin photoadducts in mouse fibroblasts: Mutations at cross-linkable sites induced by monoadducts as well as cross-links. Cancer Res (1995) 55: 1283–1288. | PubMed | ISI | ChemPort |
13.	Kanjilal S, Strom S & Clayman G et al. p53 mutations in nonmelanoma skin cancer of the head and neck: Molecular evidence for field cancerization. Cancer Res (1995) 55: 3604–3609. | PubMed | ISI | ChemPort |
14.	Kreimer-Erlacher H, Seidl H, Bäck B, Kerl H & Wolf P. High mutation frequency at Ha-ras exons 1-4 in squamous cell carcinomas from PUVA-treated psoriasis patients. Photochem Photobiol in press.
15.	Kripke ML, Morison WL & Parrish JA. Induction and transplantation of murine skin cancers induced by methoxalen plus ultraviolet (320–400 nm) radiation. J Natl Cancer Inst (1982) 68: 685–690. | PubMed | ChemPort |
16.	Maier H, Schemper M, Ortel B, Binder M, Tanew A & Hönigsmann H. Skin tumors in photochemotherapy for psoriasis: a single-center follow-up of 496 patients. Dermatology (1996) 193: 185–191. | PubMed | ISI | ChemPort |
17.	McKenna KE, Patterson CC, Handley J, McGinn S & Allen G. Cutaneous neoplasia following PUVA therapy for psoriasis. Br J Dermatol (1996) 134: 639–642. | PubMed | ISI | ChemPort |
18.	Moles JP, Moyret C, Guillot B, Jeanteur P, Guilhou JJ, Theillet C & Basset-Seguin N. p53 gene mutations in human epithelial skin cancers. Oncogene (1993) 8: 583–588. | PubMed | ISI | ChemPort |
19.	Nataraj AJ, Trent JC, II & Ananthaswamy HN. p53 gene mutations and photocarcinogenesis. Photochem Photobiol (1995) 62: 218–230. | PubMed | ISI | ChemPort |
20.	Nataraj AJ, Black HS & Ananthaswamy HN. Signature p53 mutation at DNA cross-linking sites in 8-methoxypsoralen and ultraviolet A (PUVA) -induced murine skin cancers. Proc Natl Acad Sci USA (1996) 93: 7961–7965. | Article | PubMed | ChemPort |
21.	Nataraj AJ, Wolf P, Cerroni L & Ananthaswamy HN. p53 mutations in squamous cell carcinomas from psoriasis patients treated with psoralen + UVA (PUVA). J Invest Dermatol (1997) 109: 238–243. | Article | PubMed | ISI | ChemPort |
22.	Olschwang S, Laurent-Puig P, Vassal A, Salmon RJ & Thomas G. Characterization of a frequent polymorphism in the coding sequence of the Tp53 gene in colonic cancer patients and a control population. Hum Genet (1991) 86: 369–370. | PubMed | ISI | ChemPort |
23.	Peritz AE & Gasparro FP. Psoriasis, PUVA, and skin cancer3/4molecular epidemiology: The curious question of TA transversions. J Invest Dermatol Proc The (1999) 4: 11–16. | ChemPort |
24.	Ponten F, Berg C, Ahmadian A, Ren ZP, Nister M, Lundeberg J, Uhlen M & Ponten J. Molecular pathology in basal cell cancer with p53 as a genetic marker. Oncogene (1997) 15: 1059–1067. | PubMed | ISI | ChemPort |
25.	Rady P, Scinicariello F, Wagner RFJ & Tyring SK. p53 mutations in basal cell carcinomas. Cancer Res (1992) 52: 3804–3806. | PubMed | ISI | ChemPort |
26.	Reid T & Loeb L. Tandem double CCTT mutations are produced by reactive oxygen species. Proc Natl Acad Sci USA (1993) 90: 3905–3907.
27.	Sage E, Drobetsky EA & Moustacchi E. 8-Methoxypsoralen-induced mutations are highly targeted at crosslinkable sites of photoaddition on the non-transcribed strand of a mammalian chromosomal gene. EMBO J (1993) 12: 397–402. | PubMed | ISI | ChemPort |
28.	Sato M, Nishigori C, Zghal M, Yagi T & Takebe H. Ultraviolet-specific mutations in p53 gene in skin tumors in xeroderma pigmentosum patients. Cancer Res (1993) 53: 2944–2946. | PubMed | ISI | ChemPort |
29.	Shibutani S, Takeshita M & Grollmann AP. Insertion of specific bases during DNA synthesis past the oxidation-damaged base 8-oxodG. Nature (1991) 349: 431–434. | Article | PubMed | ISI | ChemPort |
30.	Stern RS, Nichols KT & Vakeva LH. Malignant melanoma in patients treated for psoriasis with methoxsalen (psoralen) and ultraviolet A radiation (PUVA). The PUVA Follow-up Study. N Engl J Med (1997) 336: 1041–1045. | Article | PubMed | ISI | ChemPort |
31.	Stern RS, Liebman EJ & Vakeva L. Oral psoralen and ultraviolet-A light (PUVA) treatment of psoriasis and persistent risk of nonmelanoma skin cancer. The PUVA Follow-up Study. J Natl Cancer Inst (1998) 90: 1278–1284 10.1093/jnci/90.17.1278. | Article | PubMed | ChemPort |
32.	Strauss GH, Bridges BA, Greaves M, Hall-Smith P, Price M & Vella-Briffa D. Inhibition of delayed hypersensitivity reaction in skin (DNCB test) by 8-methoxypsoralen photochemotherapy: Possible basis for pseudo-promoting action in skin carcinogenesis? Lancet (1980) ii: 556–559.
33.	Wang XM, McNiff JM, Klump V, Asgari M & Gasparro FP. An unexpected spectrum of p53 mutations from squamous cell carcinomas in psoriasis patients treated with PUVA. Photochem Photobiol (1997) 66: 294–299. | PubMed | ISI | ChemPort |
34.	Wang D, Kreutzer DA & Essigmann JM. Mutagenicity and repair of oxidative DNA damage: Insights from studies using defined lesions. Mutat Res (1998) 400: 99–115 10.1016/s0027-5107(98)00066-9. | Article | PubMed | ISI | ChemPort |
35.	Wikonkal NM & Brash DE. Ultraviolet radiation induced signature mutations in photocarcinogenesis. J Invest Dermatol Symp Proc The (1999) 4: 6–10. | ChemPort |
36.	Ziegler A, Leffell DJ, Kunala S & Sharma HW. Mutation hotspots due to sunlight in the p53 gene of nonmelanoma skin cancers. Proc Natl Acad Sci USA (1993) 90: 4216–4220. | PubMed | ChemPort |

Acknowledgments

We thank Dr. Renate Schöllnast for help in collecting clinical data. This work was supported by grant no. 12383-GEN from the Austrian Science Fund (FWF). H. K.-E. was supported by the Austrian National Bank Jubilee Fund no. 7285.